Flux-sum analysis identifies metabolite targets for strain improvement.

Meiyappan Lakshmanan,Tae Yong Kim,Bevan K S Chung,Dong-Yup Lee,Sang Yup Lee

doi:10.1186/s12918-015-0198-3

Abstract

BackgroundRational design of microbial strains for enhanced cellular physiology through in silico analysis has been reported in many metabolic engineering studies. Such in silico techniques typically involve the analysis of a metabolic model describing the metabolic and physiological states under various perturbed conditions, thereby identifying genetic targets to be manipulated for strain improvement. More often than not, the activation/inhibition of multiple reactions is necessary to produce a predicted change for improvement of cellular properties or states. However, as it is more computationally cumbersome to simulate all possible combinations of reaction perturbations, it is desirable to consider alternative techniques for identifying such metabolic engineering targets.ResultsIn this study, we present the modified version of previously developed metabolite-centric approach, also known as flux-sum analysis (FSA), for identifying metabolic engineering targets. Utility of FSA was demonstrated by applying it to Escherichia coli, as case studies, for enhancing ethanol and succinate production, and reducing acetate formation. Interestingly, most of the identified metabolites correspond to gene targets that have been experimentally validated in previous works on E. coli strain improvement. A notable example is that pyruvate, the metabolite target for enhancing succinate production, was found to be associated with multiple reaction targets that were only identifiable through more computationally expensive means. In addition, detailed analysis of the flux-sum perturbed conditions also provided valuable insights into how previous metabolic engineering strategies have been successful in enhancing cellular physiology.ConclusionsThe application of FSA under the flux balance framework can identify novel metabolic engineering targets from the metabolite-centric perspective. Therefore, the current approach opens up a new research avenue for rational design and engineering of industrial microbes in the field of systems metabolic engineering.Electronic supplementary materialThe online version of this article (doi:10.1186/s12918-015-0198-3) contains supplementary material, which is available to authorized users.

Highlights

Rational design of microbial strains for enhanced cellular physiology through in silico analysis has been reported in many metabolic engineering studies
First, the conventional constraints-based flux analysis problem is solved with biomass maximization as objective and the wild-type flux-sum of all metabolites are calculated as a reference
To demonstrate the applicability of this proposed framework, we apply it to the E. coli genomescale metabolic model, thereby identifying the possible metabolite attenuation/intensification targets which can enhance the production of ethanol/succinate, and reduce the formation of toxic acetate

Summary

Introduction

Rational design of microbial strains for enhanced cellular physiology through in silico analysis has been reported in many metabolic engineering studies. The successful application of GEM analysis to aid cellular metabolic engineering has been consistently reported in numerous studies ranging from simple in silico simulation of gene deletions [10, 11], to more sophisticated computational techniques such as OptKnock [12], OptReg [13], OptStrain [14], OptGene [15], flux response analysis [16], RobustKnock [17], flux scanning based on enforced objective flux [18], OptForce [19] and most recently, cofactor modification analysis [20] for identifying valid gene knockout, up-, downregulation and cofactor engineering targets (see [21] for thorough review) These in silico methods share the common theme of identifying suitable genetically and environmentally perturbed conditions by optimizing the cellular objective, typically cell growth, under mass balance, reaction reversibility and flux capacity constraints. To deal with this limitation, an alternative method is needed to further expand our in silico capability of multiple reaction targets identification

Methods

Results

Conclusion